Research Projects

This summary is divided into four parts:

  1. Basic laboratory research interest
  2. Past studies
  3. Current projects (as of March 2009)
  4. Why the world might actually care about these things

Basic Laboratory Research Interests


My laboratory is interested in how mammalian cells control actin filament assembly. By my count, actin filaments are crucial to at least 15 distinct cellular processes in mammals (Figure 1), and I am certain there are more processes to be found. All of these structures use the same helical actin filaments, but often organize these filaments into very different higher order assemblies to conduct their specific functions. The actin monomers for all of these structures come from a common cytosolic pool, and are dynamically cycling between structures.

To put it another way, actin monomers are like bricks, that can be used for many purposes depending on what is needed. For instance, they can be made into a garden path at times, a patio at others, or a wall on occasion. None of these structures are permanent, and the bricks are constantly flitting in and out, round about, going from wall-to-brick pile-to-patio-to-brick pile-to-path-to-brick pile, and back again. Think of the world Alice encounters through the looking glass, and you won't be far off.

In the face of all this, how does a cell trigger the assembly of a particular structure exactly where and when it is needed? "To be or not to be?" isn't the only question. Others are "When?",
"Where?", and "For how long?".

One answer to this question is "assembly factors". These are proteins that accelerate actin filament assembly. One major family of actin assembly factors is the formin protein family. In mammals, there are 15 distinct formin proteins. In principle, assembly of many of these actin-containing structures could be controlled by distinct formins. In practice, it's not as simple as that, but the large number of formins does provide the possibility for differential regulation. See reviews from my laboratory (Nicholson-Dykstra 2005, Chhabra & Higgs 2008) for more details.

In my laboratory, we start with the biochemical characteristics of formin proteins, and build towards an understanding of their cellular functions.

Past Studies

Our early work was to figure out the basic biochemical features of formins. We focused on the formin FH2 domain, which is common to all formins. The FH2 is dimeric, and assembles into a donut-like structure. The FH2 alone affects actin assembly by:

These basic properties are described in several of our papers (Li 2003, Harris 2004, Harris 2006, Kovar 2006) and illustrated in Figure 2.

We have conducted other biochemical studies on formin protein regulation (Li 2003, Li 2005, as well as phylogenetic work on formins (Higgs & Peterson 2005). Other past work, related to our current projects, is described below.

Current Projects

We have three current projects, building from initial biochemical studies to investigate formin function in cells.

INF2: a polymerization/depolymerization factor that binds ER and mitochondria
INF2 is a unique formin, both biochemically and in its cellular properties. In fact, routinely has us pulling at our hair in frustration, until we figure out that what we're looking at is unique and cool. When we started working on INF2, it was simply an entry in the DNA sequence database (and was mis-annotated at that!). We have been pulling hair/learning about it since 2003.

Biochemically, INF2 is the only formin that can both accelerate actin polymerization and actin depolymerization. We figured out the basic mechanism of how this work (Chhabra 2006). Figure 3 shows some of the data, for the aficionados. INF2 has a unique C-terminus. The combination of the FH2 and the C-terminus is a very potent severing protein, chopping filaments up into smaller fragments. Severing only occurs when the actin filament is in its ADP form, providing the timer for depolymerization to occur. Actin monomers are ATP-bound. When an actin monomer adds to a filament, it hydrolyzes its ATP, to make ADP and inorganic phosphate (Pi). Both ADP and Pi remain bound to the actin subunit in the filament for some time. Eventually, the Pi releases, resulting in the actin subunit becoming the ADP form.

In addition, the C-terminus contains an actin monomer-binding WH2 motif. Mutation of the WH2 blocks depolymerization, but not the severing activity. So, the WH2 conducts a second step in depolymerization, after severing.

Our working model for polymerization/depolymerization is shown in Figure 4. We have MANY, MANY, MANY more biochemical questions to answer concerning this mechanism! However, we decided it was time to work out what it's doing in cells, resulting in more hair pulling/learning.

In cells, much of the INF2 is tightly bound to intracellular membranes (Chhabra 2009 in press). The membrane binding is dependent on a post-translational modification, prenylation, which occurs at the C-terminus. However, this is not the only thing required to keep INF2 stuck to membranes. We would like to know what the other components are.


So, what are these intracellular membranes to which INF2 is bound? More hair pulling! In mouse fibroblasts, INF2 is confined to the endoplasmic reticulum (ER, Figure 5). In other cells, INF2 appears to be associated with mitochondria (Figure 6)! However, we believe that this "mitochondrial" localization might really be localization to specialized ER that associates with mitochondria. We are actively working on figuring this out.

So, what does INF2 do on ER, mitochondria, or wherever it's found? What we know now is that cells subjected to RNA interference for INF2 get very sick, very quickly, and die. Difficult to study, but we are breaking this down now.

The most pressing questions we'd like to address for INF2 are:

  1. How does it decide what membranes to bind? Specific binding proteins?
  2. What is its cellular function on these membranes? Does it act in apoptotic pathways? That's one of our leading theories.
  3. How does its biochemical activity (actin polymerization) mediate its cellular activity?

Question 3 will involve some fascinating biochemical studies. We now have purified INF2 that we expressed in insect cells, and that contains the prenyl modification. We would like to do some in vitro microscopy on its actin polymerization/depolymerization activity, using Total Internal Reflection (TIRF). If we bind INF2 to model membranes, do we see active polymerization/depolymerization on the membrane surface? If so, can we distinguish whether the filaments produced are extremely short, with the same INF2 molecule polymerizing at one end (via the FH2) and depolymerizing at the other (via the WH2)? These filaments would be treadmilling extremely rapidly, which would make the filament similar to a conveyor belt at the grocery store. Why would this be happening in the cell? Alternately, one INF2 molecule might assemble a filament, and another one might start depolymerizing it on the ER surface. This would give longer filaments. Why? These would be fun studies.

FRL2: a bundling formin that causes assembly of microvilli
We found that a sub-set of mammalian formins can also bundle actin filaments (Harris 2006). These formins include mDia2, FRL1, and FRL2. The bundling activity requires only the FH2 domain. We came up with a really exciting possible biochemical mechanism for the bundling activity, in which the FH2 dimer is able to dissociate, then re-associate around the filament (Figure 7). This biochemical model requires much more testing!

We have focused on one bundling formin, FRL2, because of its potent ability to induce "finger-like protrusions", either filopodia or microvilli, in cells. I show one example (Figure 8), but this happens in many cell types. You can see that FRL2 concentrates at the tips of these microvilli. FRL2 can even convert other types of actin-based structure to microvilli in some cell types. Microvilli assembly by FRL2 requires the FH1 and FH2 domains. Amazingly, similar FH1-FH2 constructs of the closely related FRL1 protein do not have this activity.

We have several questions here, all wide open for investigation:

  1. What makes FRL2 such a potent microvilli-inducing protein, whereas FRL1 is not? We would like to do domain-swap experiments and point mutagenesis to find out what the key regions of FRL2 are for this activity.
  2. To what does FRL2 bind on the plasma membrane? To assemble microvilli, FRL2 must interact with the plasma membrane, either directly or indirectly. We would like to look for binding partners for this interaction.
  3. What is FRL2 doing in cells? Why is generation of microvilli important? We would like to address this by RNA interference.

Chemical inhibitors for formin proteins
In collaboration with Jeff Peterson (Fox Chase Cancer Center), we are working on chemical inhibitors for specific formin proteins. These should be valuable tools in assessing both biochemical mechanism and cellular function. With a lot of luck, they could also turn into therapies.

Why the world might actually care

The ability to control individual actin assembly processes provides the potential to mediate a large number of processes important in health and disease. Here are three examples, but the as-yet-to-be-identified functions of many of the formin proteins will surely provide new therapeutic targets.